SIERRA NEVADA PERIJÁ

Many things motivate human beings: the need for food, the need for shelter, the need to eat ice cream while watching late-night television (obviously some needs are more fundamental than others). After primary needs such as food and shelter are satisfied, the need to belong is among the strongest of human motivations (Baumeister & Leary, 1995). People go to extraordinary lengths to affiliate with others, be liked by others, and belong to groups. These needs might have arisen from evolutionary pressures; our ancestors who were excluded from social groups often died because they found it difficult to hunt, gather, and defend themselves against predators with only an army of one (e.g., Ainsworth, 1989; Axelrod & Hamilton, 1981; Barash, 1977; Bowlby, 1969; Buss, 1990, 1991; Hogan, Jones, & Cheek, 1985; Moreland, 1987). In addition, people excluded from groups were, almost by definition, unlikely to reproduce themselves. Thus the solitary hunters we described in the previous sentence were, most likely, no one’s ancestors, even if they did manage to live out a normal lifespan.

Mark Leary and his colleagues have proposed that low self-esteem often results from feelings of rejection and loneliness (e.g., Leary & Baumeister, 2000; Leary, Tambor, Terdal, & Downs, 1995; Leary & Downs, 1995). Thus when people feel disliked by others, they often internalize these feelings and come to feel dislike for themselves. Leary suggests that self-esteem acts as a “sociometer” that measures a person’s prospects for belongingness. High self- esteem means that one is the sort of person with whom others will want to affil- iate, and low self-esteem means that one is the sort of person who may be neglected or even shunned by others. Because the need to belong is a power- ful motivator, self-esteem acts as a “gas gauge” for our “fuel tank” of stored belongingness; people will make every effort to replenish it if it gets too low.

In our research, we sought to expand upon this previous research, address- ing the behavioral consequences of social exclusion. When people are rejected by others, how do they react? When people hear they are likely to be alone later in life, what patterns can we find in their subsequent behavior?

Aggression: If You Can’t Join Them, Beat Them

Over the past few years, Americans have been shocked and saddened by a series of shootings at our nation’s high schools. In most of these cases, the

perpetrators have been young men who felt rejected and bullied by their peers (Leary, Kowalski, Smith, & Phillips, 2003). The example of school violence is not an isolated one: social rejection or exclusion often leads to impulsive acts. Children who are rejected by their peers are more aggressive (Coie, 1990; Newcomb, Bukowski, & Pattee, 1993), and many perpetrators of violence are young men who feel cut off from close relationships with family members as well as with age mates (Garbarino, 1999; Walsh, Beyer, & Petee, 1987).

Adults are not immune to this effect. For example, single men commit more crimes than married men do, even when age is controlled (Sampson & Laub, 1993). They are more likely to be arrested for speeding or reckless driv- ing (Harrington & McBride, 1970), and are more likely to be involved in car accidents (Harano, Peck, & McBride, 1975), especially those related to alcohol (Richman, 1985). In general, single people abuse alcohol and drugs at a higher rate than married people (Williams, Takeuchi, & Adair, 1992). These trends also appear in society at large. In the United States, social bonds between peo- ple have weakened in recent decades (Fukuyama, 1999; Putnam, 2000) at the same time that impulsive antisocial behaviors such as crime, drug abuse, and violence have increased.

All of these sources and trends suggest that social exclusion is correlated with increased aggressive and antisocial behavior. However, the opposite rela- tion is actually more logical: when one is rejected, it makes sense to be less aggressive and more prosocial, in an attempt to win back friends and establish affiliation. In fact, Williams, Cheung, and Choi (2000) found that participants who were ostracized were later more likely to conform to other’s judgments, which those researchers regarded as prosocial act designed to make the person more appealing to potential group members. Whether social exclusion is related to more aggressive or less aggressive behavior, the previous research gives little suggestion of the direction of causation. For example, social exclu- sion could cause aggression, but it is equally plausible that aggressive behavior could cause exclusion by others. Mediation by negative emotion might also occur. Given these unanswered questions, we undertook a series of experimen- tal studies to address the question of social exclusion and aggression (see Table 2.1).

The most straightforward tests of the link between social exclusion and aggression occurred in Experiments 4 and 5 of Twenge, Baumeister, Tice, and Stucke (2001). In Experiment 4, participants spent the first 15 minutes of the study interacting with a group of their same-sex peers. They were then placed in separate rooms and asked to nominate the two group members they wanted to work with on a subsequent task. By random assignment, half of the partici- pants then heard that none of the other participants had chosen them (this was the rejected condition). The other half heard that everyone had chosen them (accepted condition). All participants were then provoked: they received a very negative evaluation of an essay that they had written, supposedly issued by a

new participant who had just arrived at the lab (and thus not one of the people who had accepted or rejected them). Participants then played a noise-blasting game with this new person; they were told that whoever lost a reaction time trial would hear a blast of noise, the intensity and duration of which was con- trolled by the other player. Thus participants had a weapon that they could use to hurt another human being. We wanted to know if rejected participants would be more or less aggressive than accepted participants.

The results clearly showed that rejected participants were more aggressive toward the person who insulted them. Rejected participants blasted louder noise at their opponents and were also more aggressive on a composite meas- ure of aggression (consisting of noise loudness and duration). The two condi- tions did not differ on a self-report measure of emotion, and the results linking social exclusion and aggression did not change when controlled for negative or positive emotion.

In the next study, we sought to broaden the focus. The preceding study showed that rejected people became highly aggressive toward someone who insulted and provoked them. Would they also be aggressive toward someone who had not insulted and provoked them? We used the same procedure save one important change: in this study, the other person did not deliver any critical insult- ing evaluation to the participants. They did not receive any evaluation of their essay from the opponent; in fact, they had no interaction at all with the person with whom they played the noise-blasting game. The results showed that rejected par- ticipants were still significantly more aggressive than participants who had been accepted: they blasted louder noise and for a longer time. Thus social exclusion causes an increase in aggressive behavior even toward an innocent target. TABLE 2.1 Summary of Results of Laboratory Studies of Social Exclusion

Project/dependent variable Effect of exclusion

Aggression (Twenge, Baumeister, Tice, & Blasting unpleasant noise, negative job evaluation. Stucke, 2001; Twenge & Cacho, 2003) Only occurs in some situations.

Prosocial behavior (Twenge, Ciarocco, Donating less money, volunteering less time, not Cuervo, Bartels, & Baumeister, 2004) helping experimenter, not cooperating in game Self-defeating behavior (Twenge, Risky lottery choice, choice of unhealthy behaviors,

Catanese, & Baumeister, 2002) procrastination

Cognitive performance (Baumeister, Poor performance on: IQ test, GRE reading Twenge, & Nuss, 2002) comprehension, GRE analytical reasoning Individual differences: Narcissism and Narcissism leads to aggression; self-esteem

self-esteem (Twenge & Campbell, 2003) has no effect

Mood and emotion (Twenge, Catanese, & No effects or very weak effects on explicit measures; Baumeister, 2003; Twenge & less emotion on implicit measures

We obtained converging evidence for these conclusions by using a different manipulation of social exclusion and a different measure of aggression. In this experiment, we manipulated social exclusion by giving participants false feed- back on a personality test. In the crucial condition (future alone), participants heard that they were likely to be alone later in life. In contrast, participants in the future belonging condition heard that they would have good relationships throughout life. We also included a misfortune control group, who heard that they would likely be accident prone in the future (thus a negative outcome, but one unrelated to relationships). After receiving the future prediction, all participants were then provoked by a negative essay evaluation. They were told that the person who had evaluated their essay had applied for a research assis- tant position that was very competitive; the participant was asked to evaluate the applicant, rating him/her on a list of attributes. Thus, the participant could hurt or help someone who had insulted them. Written evaluations have been used to measure aggression in several other studies (e.g., Kulik & Brown, 1979; Ohbuchi, Kameda, & Agarie, 1989; O’Neal & Taylor, 1989; for a review, see Baron & Richardson, 1994, pp. 64–66).

Consistent with the results of the noise-blasting studies, participants who heard that they were likely to be alone later in life were more aggressive toward the person who had insulted them, issuing a sharply negative job evaluation. In contrast, the future belonging and misfortune control groups gave neutral eval- uations of the applicant. The difference with the misfortune control group is informative; the forecast of a negative future is apparently not the cause of the heightened aggression. Only the forecast of a lonely future causes a notable increase in aggression.

Last, we sought to establish whether excluded individuals would also be aggressive toward someone who praised them. We used the same procedure as in the previous study, except participants received a positive evaluation of their essay. Under these circumstances, we found that all participants gave a positive evaluation of the applicant, with no differences between the future alone, future belonging, and misfortune control. Thus socially excluded participants are more aggressive than others when they are insulted and when they are neither insulted nor praised; however, they are not more aggressive toward someone who praises them.

Thus, this series of experiments shows that social exclusion can cause a marked increase in aggressive behavior. The average effect size for this set of studies is 1.33, meaning that rejected individuals scored 1.33 standard deviations higher on measures of aggression compared to the control groups. Anything over 0.80 standard deviations is conventionally regarded as a large effect, so the impact of social exclusion appears to be quite large. Several other researchers have repli- cated or partially replicated these results, also finding that social exclusion leads to increased aggression (Buckley, Winkel, & Leary, 2004; Kirkpatrick, Waugh, Valencia, & Webster, 2002; Warburton, Williams, & Cairns, 2003).

Why does social exclusion lead to increased aggression? Negative emotional states do not seem to be the culprit, as they did not mediate the effect in any of the studies. Although definitive evidence is not available, it seems plausible that socially excluded individuals become antisocial because they no longer see the point in being prosocial. When we are accepted, our social behavior is shaped by the demands of others; when we are rejected, we may indulge our more selfish and base impulses and act more aggressively.

We have begun to explore which situations cause aggression after rejection, and which can prevent it (Twenge & Cacho, 2003). In one experiment, half of the participants heard that they would interact with their game partner later, and the other half heard that they would not (the usual situation). When rejected participants heard that they would interact with their game partner, they were no longer aggressive. In the second study, rejected and accepted par- ticipants were told that their game partners were accepted by another group, rejected by another group, or simply members of another group (neutral). Rejected participants were not aggressive toward neutral group members, but were aggressive toward both accepted and rejected targets. Thus aggression after rejection is heightened when participants do not expect to interact further with the target of their aggression, and when they believe that the target is either different from them (accepted) or low status (rejected). On the other hand, aggression after rejection does not occur when participants expect to interact with their game partner, and when they believe that the target is from another group. This is consistent with the motivational theory of belongingness (Baumeister & Leary, 1995), which predicts that rejected people should seek acceptance in order to regain social ties. These results show that rejected people will act prosocially when they interact with those who may provide social acceptance (a future interaction partner, or another group member who is not too high or too low in status).

Prosocial behavior: Help! I Need Somebody

Although we find that social exclusion causes aggressive behavior, it could be that excluded individuals could act in prosocial ways when they are given the explicit opportunity to do so. Thus we undertook another series of experiments to explore a possible link between social exclusion and prosocial behavior. Previous evidence and theory suggests that social exclusion could cause either increased or decreased prosocial behavior.

If people are strongly motivated to form and maintain social bonds (Baumeister & Leary, 1995), then the loss of social connection frustrates this basic human need. This should motivate people to desire new connections. By helping others and cooperating with them, people could presumably prove their social value to others; in this way they could induce others to like and depend on them. It seems only rational that the socially excluded person would

try harder to get along with others, and an increase in prosocial behavior seems a promising way to accomplish this. There is some empirical evidence to support this idea (e.g., Gardner, Pickett, & Brewer, 2000; Williams et al., 2000; Williams & Sommer, 1997).

The opposite prediction, that social exclusion could reduce prosocial behavior, is also plausible. Under this view, prosocial behavior depends on believing one is part of a community in which people mutually seek to aid, sup- port, and love each other. When people are excluded, their motivation to per- form such behaviors would be reduced or eliminated. Excluded people might feel that following social rules is no longer necessary, or that there is no point in continuing to make any such efforts or sacrifices. In many cases, doing what is prosocial and right for the group conflicts with self-interest. Most concep- tions of virtue and socially desirable behavior promote prosocial actions that go against the individual’s own wishes and desires (e.g., Hogan, 1973; Baumeister & Exline, 1999). For example, children must be socialized into helping others (Cialdini & Kenrick, 1976; Cialdini, Kendrick, & Baumann, 1981; Perry, Perry, & Weiss, 1986). Freud (1930) proposed that the superego (an internal system of restraint) is crucial for making civilized communal life possible: It emerges as a capacity to thwart instinctual or selfish interests in order to pursue actions that are valued by the group. Without the socializing context provided by social inclusion, the psychological restraints that support prosocial behavior may be diminished, and the prosocial impulse might be extinguished.

Prior research also provides correlational support for a connection between social rejection and decreased prosocial behavior, although it is unclear what causes what. Numerous correlational studies have found that children who are rejected by their peers act less prosocially than children who are accepted by their peers (Asher & Coie, 1990; Coie & Dodge, 1988; Gest, Graham-Berman, & Hartup, 2001; Mize & Ladd, 1988; Wentzel & Erdley, 1993; Wentzel & McNamara, 1999). Many studies have found that prosocial actions are more highly correlated with social acceptance than antisocial actions are (Asher & Renshaw, 1981; Bukowski & Newcomb, 1984; Hartup, Glazer & Charlesworth, 1967; Parkhurst & Asher, 1992; Schonert-Reichl, 1999). However, these find- ings are correlational, so it is unclear if social exclusion causes less prosocial behavior or vice versa.

We performed a series of experimental studies to determine if social exclu- sion causes more or less prosocial behavior (Twenge et al., 2004). In the first experiment, we used the future outcomes manipulation mentioned previously: future alone participants heard that they were likely to be alone later in life, compared to three control groups (future belonging, misfortune control, and a pure control group hearing no future prediction). After the manipulation, the experimenter gave each participant two dollars in quarters. She then mentioned that the laboratory was taking up a collection for the Student Emergency Fund, and pointed to a collection box on the table. The amount of

money donated by each participant served as the measure of prosocial behavior. The results were striking: on average, future alone participants donated less than $0.40, while participants in the other three conditions donated an average of $1.50—that is, nearly three times as much. Only 37% of the Future Alone participants made any donation at all, whereas every single participant (100%) in the other three conditions gave at least something. These results were not mediated by either self-reported emotion or by state self-esteem.

In the second experiment, we manipulated social exclusion via acceptance or rejection by peers. The experimenter said that the participant could not complete the regular part of the experiment, so he or she could either leave or could help the experimenters by doing between one and three short studies for the remainder of the hour. The experimenter explained that the experimental credit would be the same either way. The results of this experiment again showed that excluded participants were markedly less helpful. Rejected partic- ipants volunteered for only 0.30 extra experiments, whereas accepted partici- pants volunteered for 1.70 experiments on average. Only 20% of rejected people volunteered to help at all, whereas 90% of accepted people volunteered to help with at least one study.

The first two experiments both measured prosocial behavior that involved the sacrifice of self-interest: participants were asked to give up a scarce resource (either money or time). In the third experiment, we measured a prosocial behavior that did not involve a noticeable sacrifice. We manipulated social exclusion using the prediction of future outcomes. After the participant completed an emotion measure, the experimenter reached toward a shelf and knocked over a can of pencils. The measure of helping was the number of pen- cils the participant helped pick up from the floor. Even though this form of helping did not involve a sacrifice, socially excluded participants were still con- siderably less helpful compared to the people in the other groups. The future alone group barely helped at all; on average, they helped pick up less than one pencil. In contrast, the other groups helped pick up between eight and nine pencils on average. Only 15% of the Future Alone participants helped pick up any pencils at all, compared to the 64% who helped in the other three condi-